Quick side example. You are a hit capped warlock with 30% crit. You add a new trinket that adds 2% crit. By the above argument, 98% of the time that trinket is doing absolutely nothing for you (30% of the time you would have crit anyway, and 68% of the time you would have hit anyway). Does that make this new 2% crit trinket useless? No, of course not; 2% of the time it will make a hit into a crit.
When Bladeward procs and you get the buff, you get 2.5% avoidance for 10 seconds (just taking a round number since it of course varies depending on your current parry and defense ratings, in my gear a 1-stack would be 2.58% parry after DR). Now regardless of if you "use" that 2.5% or not is irrelevent; you increased your avoidance by 2.5% for 10 seconds, so 2.5% of the time something that would have hit is now instead avoided.
Likewise for Mongoose, 120 extra agility (at my gear level) is an additional 1.07% dodge after DR for 15 seconds. Again, regardless of any specific combat roll in which I might have dodged anyway, I still gained that 1.07% avoidance. 99%, by some people's arguments, I'm not actually "using" the Mongoose proc and it is doing nothing for me, but it doesn't change the fact that it increased my avoidance by 1.07% for those 15 seconds.
So therefore, in order to figure out how much avoidance each enchant is providing me in the grand scheme of things, we need to average out the avoidance gain by the "uptime" of the buff. Unfortunately Bladeward's buff is consumed if you parry while it is up, which has the effect of lowering Bladeward's uptime. This makes it really hard to determine the overall avoidance gains of this enchant and really is what makes it lackluster. IF bladeward lasted it's full 10 second duration and had a 20% uptime, it would provide an average of 0.5% constant avoidance. However, there's absolutely no way of telling how much uptime (in theory) you are going to have on the buff because of the random nature of the buff being consumed.
Mongoose is much more straight forward. Depending on what numbers you look at, Mongoose has around 40% uptime. So 40% of the time you gain 1.07% avoidance for an average 0.428 "constant" avoidance.
Personally, I don't like either of them. I'll take my constant 25 hit rating and constant 25 crit rating from Accuracy...